transformers/docs/source/en/model_doc/marian.md

MarianMT

Overview

MarianMT is a machine translation model trained with the Marian framework which is written in pure C++. The framework includes its own custom auto-differentiation engine and efficient meta-algorithms to train encoder-decoder models like BART.

All MarianMT models are transformer encoder-decoders with 6 layers in each component, use static sinusoidal positional embeddings, don't have a layernorm embedding, and the model starts generating with the prefix pad_token_id instead of <s/>.

You can find all the original MarianMT checkpoints under the Language Technology Research Group at the University of Helsinki organization.

Tip

This model was contributed by sshleifer.

Click on the MarianMT models in the right sidebar for more examples of how to apply MarianMT to translation tasks.

The example below demonstrates how to translate text using [Pipeline] or the [AutoModel] class.

import torch
from transformers import pipeline

pipeline = pipeline("translation_en_to_de", model="Helsinki-NLP/opus-mt-en-de", torch_dtype=torch.float16, device=0)
pipeline("Hello, how are you?")

import torch
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-en-de")
model = AutoModelForSeq2SeqLM.from_pretrained("Helsinki-NLP/opus-mt-en-de", torch_dtype=torch.float16, attn_implementation="sdpa", device_map="auto")

inputs = tokenizer("Hello, how are you?", return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, cache_implementation="static")
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Use the AttentionMaskVisualizer to better understand what tokens the model can and cannot attend to.

from transformers.utils.attention_visualizer import AttentionMaskVisualizer

visualizer = AttentionMaskVisualizer("Helsinki-NLP/opus-mt-en-de")
visualizer("Hello, how are you?")

Supported Languages

All models follow the naming convention: Helsinki-NLP/opus-mt-{src}-{tgt}, where src is the source language code and tgt is the target language code.

The list of supported languages and codes is available in each model card.

Some models are multilingual; for example, opus-mt-en-ROMANCE translates English to multiple Romance languages (French, Spanish, Portuguese, etc.).

Newer models use 3-character language codes, e.g., >>fra<< for French, >>por<< for Portuguese.

Older models use 2-character or region-specific codes like es_AR (Spanish from Argentina).

Example of translating English to multiple Romance languages:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    ">>fra<< This is a sentence in English to translate to French.",
    ">>por<< This should go to Portuguese.",
    ">>spa<< And this to Spanish."
]

model_name = "Helsinki-NLP/opus-mt-en-roa"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)

inputs = tokenizer(src_text, return_tensors="pt", padding=True)
outputs = model.generate(**inputs)
result = [tokenizer.decode(t, skip_special_tokens=True) for t in outputs]
print(result)

Notes

• MarianMT models are ~298MB on disk and there are more than 1000 models. Check this list for supported language pairs. The language codes may be inconsistent. Two digit codes can be found here while three digit codes may require further searching.

• Models that require BPE preprocessing are not supported.

• All model names use the following format: Helsinki-NLP/opus-mt-{src}-{tgt}. Language codes formatted like es_AR usually refer to the code_{region}. For example, es_AR refers to Spanish from Argentina.

• If a model can output multiple languages, prepend the desired output language to src_txt as shown below. New multilingual models from the Tatoeba-Challenge require 3 character language codes.

  add code snippet here
  
  

• Older multilingual models use 2 character language codes.

  add code snippet here